1: \begin{abstract}

2: Stochastic variance reduced gradient (SVRG) is a popular variance reduction technique for accelerating stochastic gradient descent (SGD).

3: We provide a first analysis of the method  for solving a class of linear inverse problems in the lens of the classical regularization

4: theory. We prove that for a suitable constant step size schedule, the method can achieve an optimal convergence rate in

5: terms of the noise level (under suitable regularity condition) and the variance of the SVRG iterate error is smaller than

6: that by SGD. These theoretical findings are corroborated by a set of numerical experiments.

7: 

8: \noindent\textbf{Keywords}:  stochastic variance reduced gradient; regularizing property; convergence rate; saturation; inverse problems.

9: \end{abstract}